Re: [PATCH 4/3] migration: Add tracepoints for downtime checkpoints

[Joao Martins <joao.m.martins@oracle.com>]

On 26/10/2023 20:01, Peter Xu wrote:
> Add tracepoints for major downtime checkpoints on both src and dst.  They
> share the same tracepoint with a string showing its stage.
> 
> On src, we have these checkpoints added:
> 
>   - downtime-start: right before vm stops on src
>   - vm-stopped: after vm is fully stopped
>   - iterable-saved: after all iterables saved (END sections)
>   - non-iterable-saved: after all non-iterable saved (FULL sections)
>   - downtime-stop: migration fully completed
> 
> On dst, we have these checkpoints added:
> 
>   - precopy-loadvm-completes: after loadvm all done for precopy
>   - precopy-bh-*: record BH steps to resume VM for precopy
>   - postcopy-bh-*: record BH steps to resume VM for postcopy
> 
> On dst side, we don't have a good way to trace total time consumed by
> iterable or non-iterable for now.  We can mark it by 1st time receiving a
> FULL / END section, but rather than that let's just rely on the other
> tracepoints added for vmstates to back up the information.
> 
> Until this patch, one can enable "vmstate_downtime*" and it'll enable all
> tracepoints for downtime measurements.
> 
> Since the downtime timestamp tracepoints will cover postcopy too, drop
> loadvm_postcopy_handle_run_bh() tracepoint alongside, because they service
> the same purpose, but that only for postcopy.  We then have unified prefix
> for all downtime relevant tracepoints.
> 
> Signed-off-by: Peter Xu <peterx@redhat.com>
> ---
>  migration/migration.c  | 16 +++++++++++++++-
>  migration/savevm.c     | 14 +++++++++-----
>  migration/trace-events |  2 +-
>  3 files changed, 25 insertions(+), 7 deletions(-)
> 
> diff --git a/migration/migration.c b/migration/migration.c
> index 9013c1b500..f1f1d2ae2b 100644
> --- a/migration/migration.c
> +++ b/migration/migration.c
> @@ -103,6 +103,7 @@ static int close_return_path_on_source(MigrationState *s);
>  
>  static void migration_downtime_start(MigrationState *s)
>  {
> +    trace_vmstate_downtime_timestamp("downtime-start");
>      s->downtime_start = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
>  }
>  
> @@ -117,6 +118,8 @@ static void migration_downtime_end(MigrationState *s)
>      if (!s->downtime) {
>          s->downtime = now - s->downtime_start;
>      }
> +
> +    trace_vmstate_downtime_timestamp("downtime-end");
>  }
> 

Considering we aren't including any downtime timestamps in the tracing, is this
a way to say that the tracing tool printing timestamps is what we use to extract
downtime contribution?

It might be obvious, but perhaps should be spelled out in the commit message?